Contact centers often deploy AI to cut costs. As such, they monitor its success through the lens of handling times and contact volumes.
That’s unsurprising. After all, much of the contact center AI messaging swirls around its potential to drive efficiencies. But that’s just half the story. AI also offers significant potential to make money.
Consider an automated quality assurance (QA) system. Sure, it can help spot opportunities to reduce repeat contacts, shorten interactions, and add automation. However, it may also spotlight best practices and agent behaviors that help boost customer satisfaction and loyalty.
As such, service leaders should challenge the efficiency-centric narrative surrounding contact center AI and take a broader view when measuring its success.
In doing so, they can either bolster their business case for future investment or spot potentially unintended consequences of the AI implementation, which urgently need addressing.
Ultimately, getting this right comes down to the fundamentals of contact center reporting. It requires a balanced approach to measuring critical customer, employee, and business outcomes.
Set Internal Customer, Employee, and Business Benchmarks
The key to measuring the success of AI is: how will we know success when we see it?
Think of that question through the lens of the customer, employee, and broader business. How will each critical stakeholder view the success of this AI implementation?
From there, contact centers can decipher the most relevant metrics and ensure they take a holistic approach to measuring what matters most.
Consider the implementation of a customer-facing virtual agent.
From a customer perspective, they want efficient, hassle-free service with a first-time resolution. As such, the contact center may wish to monitor metrics such as goal completion rate, missed utterances, escalation rate, return users, and customer sentiment – the same outcomes, in fact, that would be measured for a human agent.
Meanwhile, contact center employees may hope to handle more engaging and fewer repetitive conversations with a less intensive workload and less stress. Therefore, the contact center can consider tracking metrics such as agent sentiment, occupancy, and service level.
Finally, the business wants cost-efficient service that maximizes revenue. As such, consider KPIs such as total interactions, cost to serve, and overall cost of operations.
Nevertheless, Steve Nattress, VP of Product Management at Enghouse Interactive, warns:
“It’s essential to understand the nuances of each metric. Some are counterintuitive, like average handling time (AHT), which could actually go up for live agents handling more complex calls while the AHT for AI-handled calls goes down – but they still tell a story.”
After closely considering these metrics, create a picture of where the contact center is today across each. Then, define goals that reflect a broad interpretation of AI success.
But accept that achieving those goals isn’t the be-all and end-all. If the contact center misses some of these targets, it gains insight that may bolster its AI strategy over the long term.